Professional Documents
Culture Documents
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
a r t i c l e
i n f o
Article history:
Received 12 December 2008
Received in revised form 27 March 2009
Accepted 27 March 2009
Available online 7 April 2009
Keywords:
Stainless steel
Polarization
Modeling studies
Crevice corrosion
Scaling laws
a b s t r a c t
A numerical model that predicts the rates of metal dissolution and electrolyte composition along the
length of a metallic crevice was used to simulate the crevice corrosion of AISI 304 stainless steel. The
model considers both the forward and reverse electrochemical reactions that might take place during
the corrosion process. The environment exterior to the crevice, where a net cathodic current is produced,
was simultaneously modeled. It was found that cathodic reactions are likely to occur towards the tip of
the crevice. For the case when the hydrogen evolution reaction was considered as a possible cathodic
reaction in the crevice, it is shown how the delayed instigation of this reaction may be the cause of an
experimentally observed increase in pH at the crevice tip. Two critical crevice scaling laws were examined using model predictions and one scaling law t the model predictions very well. This scaling law
differentiates between crevices that will undergo active corrosion and those that will remain indenitely
passive.
2009 Elsevier Ltd. All rights reserved.
1. Introduction
Crevice corrosion is a localized form of corrosion that can cause
considerable damage. It can occur on passive metals under conditions that do not normally cause high levels of uniform corrosion.
These facts, combined with the fact that crevice corrosion can
be difcult to detect, may lead to situations where the structural
integrity of equipment is challenged by crevice corrosion under
conditions that corrosion was not expected. The damage presented
by crevice corrosion creates the need for a detailed understanding of the complex phenomena causing this type of corrosion. This
understanding would help enable the prediction and prevention of
crevice corrosion, as well as provide insight into related forms of
localized corrosion, such as stress corrosion cracking.
Crevice corrosion occurs in occluded spaces, possibly caused by
a ange or under sediment deposited on the inside of a pipe. In a
crevice a microenvironment may form in which the concentrations
of species may be considerably different from those in the bulk
electrolyte. A concentration gradient between the crevice and bulk
electrolyte can be caused by a very low rate of corrosion that manifests as a small leakage or passive current, which is always present
with passive metals. A passive current density is often much less
than 1 A/cm2 .
For metal surfaces exposed to the bulk electrolyte, the build up of
species due to the passive current is overwhelmed by mass transfer
of the species away from the surface. Therefore, the concentrations
Corresponding author. Tel.: +1 306 966 4766; fax: +1 306 966 4777.
E-mail address: Richard.Evitts@usask.ca (R.W. Evitts).
0013-4686/$ see front matter 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2009.03.080
4697
4698
2. Theory
A mechanistic model is presented that describes a metallic
crevice in various stages of corrosion. At the beginning of the simulation it is assumed that there are no concentration gradients in the
electrolyte in the crevice or in the bulk solution. The passive current that exists at the prescribed initial conditions is assumed to
exist along all metal surfaces and causes the depletion of oxygen in
the crevice. During the incubation period, changes in concentration
gradients, passive currents, and electric potentials are predicted.
If active corrosion is predicted the model determines the rate of
active corrosion within the crevice. During the propagation phase
the iR drop may cause the potential drop between anodic crevice
areas and cathodic bold areas to increase [1]. If this potential drop
becomes too great some anodic regions in the crevice will decouple
from cathodic areas on the bold surface. It is assumed (and veried
in Section 3 of this paper), that decoupled anodic regions do not
produce a net anodic current that is transported to the bold surface. This model presented here is unique as it is capable of (a)
describing the phenomena occurring during the decoupling process, and (b) possible cathodic (and anodic) reactions occurring at
decoupled regions. This includes the possibility for net cathodic
reactions occurring at decoupled regions towards the crevice tip.
2.1. Mathematical model development
Some aspects of the overall model that have been previously
published will be briey presented here for completeness [5,1218].
Kennell et al. [12] describe in detail the development of an earlier
version of the model presented in this paper. The model presented here calculates the corrosion current in an AISI 304 stainless
steel crevice by considering both the forward and reverse electrochemical reactions, instead of assuming simple Tafel kinetics.
This greatly simplies the numerical solution of the mathematical
model because a continuous function is now used to describe the
rate of substrate dissolution rather than the Tafel equation, which
was not assumed valid at potentials less than the reversible potential in the previous model. This also increases the exibility of the
model as the assumption of zero current at decoupled regions is
no longer applied [12]. This allows for an inspection analysis of
the decoupling process of anodic sites in the crevice from cathodic
sites on the bold surface. The model also includes a method for
representing the hydrogen evolution reaction (HER).
Mass transport within the crevice is determined using Eqs. (1)
and (2). Eq. (1) describes the transport of species due to convection,
diffusion, diffusion potential, and electromigration [5]:
Ci
zi ui Ci
zi ui Ci
=
i
zj Dj Cj + Di 2 Ci + S
t
F
zj2 uj Cj
zj2 uj Cj j=1
k
(1)
(2)
Table 1
Values for parameters used in simulating an AISI 304 stainless steel corroding
crevice.
Anode
i0,a
1 102pH A/cm2
E0,a
0.259 V/SHE
TS
0.03 V/decade
Ecrit
0.009 V/SHE
Cathode
i0,c [10]
1 109 A/cm2
E0,c [10]
0.191 V/SHE
TS [10]
0.1 V/decade
Ecorr
0.1 V/SHE
Cjvij Ki = 0
(3)
where Cj is the equilibrated concentration of the jth species occurring in the ith chemical reaction, and K is the equilibrium constant.
A mass balance equation can be written for each species involved
in chemical reactions [15]:
Cj = Cjinit +
n
Xi vji
(4)
i=1
The mass balance equations are combined to provide a guaranteed non-singular matrix. This matrix was solved using the
NewtonRaphson method.
Mass transport exterior to the crevice was modeled using
the equation developed by Kennell et al. [12] for an oxygenated,
convection-free electrolyte:
C
2 C
i
KMT
=D 2 +
([Bulk] [Surface])
y
zFz
t
x
(5)
Eq. (5) considers the diffusion of oxygen parallel to the bold surface, the transfer of oxygen between the bulk electrolyte and bold
surface and the consumption of oxygen at the bold surface due
to electrochemical reactions. This equation was solved implicitly
over a one-dimensional grid parallel to the bold metal surface
using central space discretization. The initial condition was bulk
concentration at all nodes and boundary conditions were zero concentration gradient at the crevice mouth and bulk concentration
at the node furthest from the crevice. The diffusion coefcient of
oxygen was assumed to be 5.5 108 m2 /s and the electrolyte was
assumed to be under atmospheric conditions.
For oxygen depleted areas in the crevice with surface overpotential greater than the critical potential the passive current was
determined by modifying a reference current for pH using a Freundlich adsorption equation [19]:
log (iP ) = log (k) npH
(6)
For all crevice areas with a surface overpotential less than the critical potential the rate of corrosion was determined by considering
both the forward and reverse reactions:
i = i0,a (10(TS (a E0,a )) 10(TS (E0,a a )) )
(7)
The critical potential, the Tafel slope, the reversible potentials and
exchange current density values used in this model are given in
Table 1. The electric potential gradient was calculated in the crevice
using [20]:
i
F
zj Dj Cj
(8)
(9)
The electric potential gradient parallel to the bold surface was calculated using [12]:
i
RT
ln
2F
[O ]
2 mouth
O2
(10)
4699
(11)
The values for the exchange current density and reversible potential
can be found in Table 1. The surface overpotential was again determined by using the potential at the crevice mouth as a reference
[12]:
c = Emouth
(12)
(13)
where the exchange current density for the hydrogen reaction was
assumed to be equal to that on iron and given by [21]:
i0,H = 1 105.9 (10pH )
0.5
(14)
and the reversible potential was calculated using the Nernst equation:
E0,H = 0.0591pH
(15)
4700
Fig. 4. Predicted electric potentials for the Alavi and Cottis [7] experimental parameters.
Fig. 2. Critical aspect ratio for crevice corrosion.
The critical aspect ratio fell between the active crevice and not
active crevice zones shown in Fig. 2. Two scaling factors from the literature were tted to this data by assuming the midpoint between
the active and not active crevice dimensions was the critical dimension for crevice corrosion. These two scaling factors were L/G [23]
and L2 /G [22]. The rst scaling factor tested, L/G, did not t the data.
The second scaling factor, L2 /G, t the data well. In fact, Fig. 2 shows
that the scaling factor L2 /G delimits the active and non-active zones
shown in the gure very well when of L2 /G = 401 cm.
Interesting observations pertaining to cathodic reactions during several of the predictions from Fig. 2 can be made. Several of
the predictions from Fig. 2 show low amounts of metal deposition
at the crevice tip during some periods of the simulation. The predicted deposition is several orders of magnitude smaller than the
amounts that would be deposited for cathodic currents of the order
of the passive current density, which equates to approximately a
total deposition of 3 ng. Fig. 3 shows the predicted net corrosion current for the crevice of 2.5 cm width, 90 10 m gap, and 8 cm depth,
which is the model validation prediction. The gure shows the
development of the net corrosion current along the crevice at different times during the simulation. As the simulation proceeds, the net
anodic currents towards the mouth of the crevice become greater
as do the concentration gradients towards the crevice mouth. This
changes the electric potential gradient and decreases the potential
at the crevice tip due to Ohmic and electrochemical considerations,
as shown in Fig. 4, which is a plot of potential distance along the
crevice. The potential inside the crevice is calculated with respect
to the bold surface. After 5 h the predicted electric potential at the
crevice tip is equal to the reversible potential of the metal at these
conditions. However, as the electric potential drop approximately
half a centimetre into the crevice mouth continues to drop due to
Fig. 3. Predicted net currents for the Alavi and Cottis [7] experimental parameters
without considering the possibility for the hydrogen evolution reaction.
4701
Fig. 7. Predicted net currents for the Alavi and Cottis [7] experimental parameters
considering the possibility for the hydrogen evolution reaction.
4702
Acknowledgments
The authors would like to thank the Natural Sciences and Engineering Research Council (NSERC) for nancial support and the
University of Saskatchewan for computing facilities.
Appendix A. Nomenclature
C
D
E0
Ecorr
Ecrit
F
i
i
io
ip
K
KMT
Q
R
S
t
T
TS
u
X
z
concentration (mol/m3 )
diffusion coefcient (m2 /s)
reversible potential (V)
corrosion potential (V)
critical potential (V)
Faradays constant (96,487 C/mol)
current density (A/m2 )
current density vector (A/m2 )
exchange current density (A/m2 )
passive current density (A/m2 )
equilibrium constant
Mass transfer coefcient (m/s)
reaction quotient
universal gas constant (8.3145 J/mol K)
source term (mol/m3 s)
time (s)
temperature (K)
Tafel slope (V/decade)
mobility (m2 mol/J s)
extent of reaction
charge number
4. Conclusions
The numerical model of crevice corrosion presented in this
paper (without considering the hydrogen evolution reaction) predicts the pH prole seen in the experimental data of Alavi and Cottis
[7]. This model demonstrates the importance of modeling both the
forward and reverse reactions of metal dissolution when modeling
localized corrosion. When the hydrogen evolution reaction was not
considered it was found that conditions may exist within a corroding crevice that may cause a net deposition of metal at the crevice
tip. This deposition would be due to an overall potential drop caused
by a combination of an iR drop coupled with a diffusion potential
drop. However, this deposition is unlikely since HER would occur
prior to metal deposition.
Simulations were used to nd a critical crevice corrosion scaling law. The scaling law that t the results was found to be L2 /G.
From data predicted by the current model it was observed that
the critical crevice geometric dimensions were described when
L2 /G 401 cm.
Due to the observed importance of cathodic reactions within a
corroding crevice after establishment of a signicant potential drop
into the crevice (due to the iR drop), further simulations were conducted that incorporated the hydrogen evolution reaction within
the crevice. It was predicted that the hydrogen evolution reaction
occurs in the crevice after concentration and potential gradients
have become signicant due to active crevice corrosion. Due to the
delay in the establishment of larger concentration and potential
gradients, the initiation of hydrogen evolution within the crevice is
delayed. This delay was responsible for an increase in the pH at the
crevice tip during the numerical simulation of the Alavi and Cottis
[7] experiment. An actual delayed increase in the pH at the crevice
tip that was recorded in the experimental data of Alavi and Cottis [7]
may be also be caused by these factors. A new mechanistic model
is also developed that illustrates some of the important factors in
crevice corrosion, including the development of a second cathodic
area towards the crevice tip.
Greek letters
electric potential
surface overpotential (V)
conductivity (1 m1 )
stoichiometric coefcient
y
distance between bulk uid and bold surface (m)
z
height of control volume (m)
Subscripts/superscripts
a
anode
c
cathode
H
hydrogen evolution reaction
i
chemical species or reaction
j
chemical species
mouth crevice mouth
last Node last node providing net cathodic current on bold surface
O2
oxygen
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[21]
[22]
[23]
[24]
[25]
4703
L.L. Shrier, Corrosion, vols. I and II, Newness Butterworth, London, 1979.
A. Turnbull, J.G.N. Thomas, J. Electrochem. Soc. 129 (1982) 1412.
Y. Xu, H.W. Pickering, J. Electrochem. Soc. 140 (1993) 658.
R.W. Evitts, Ph.D. Thesis, University of Saskatchewan, Saskatoon, SK, 1997.
S.M. Sharland, P.W. Tasker, Corros. Sci. 28 (1988) 603.